The unusually high cost of financial intermediation in Brazil, as measured by the reported difference between bank lending and deposit interest rates, is a major source of policy concern, and with good reason . Brazil is an international outlier in terms of published interest spreads—it is one of only four countries in the world that reports average spreads above 30 percentage points. The phenomenon of high interest rate margins has had a long history in Brazil, and a stubborn one at that.  While Brazil is not alone in having had very high inflation in the past and having successfully reduced it to single-digits, it is rather unique in not having achieved a reduction in interest spreads to moderate levels, as other countries with hyperinflation histories have managed to do.  This is despite an admittedly impressive improvement in fiscal and monetary performance over the past years. The high level and persistency of intermediation spreads has become an important source of concern for Brazilian policy makers and researchers, with good reason. High spreads are much more than a nuisance to the conduct of business.  They mean high, and often more volatile, lending rates, leading to a higher cost of capital, reduced investment, and a bias towards short-term high-risk investments, away from long-maturing investments with higher social returns.  Moreover, high banking spreads can disproportionately hurt small and medium enterprises and encourage informality.  More generally, high spreads can be interpreted as a symptom of a poorly functioning financial system which, of itself, can retard economic growth. Improving the functioning of the financial system is arguably is a much higher priority for development policy in Brazil than in other countries to the extent that the binding constraint to investment and growth is the shortage of finance rather than the lack of high-return investment opportunities.

While the issue of high banking spreads has received much attention, major gaps and puzzles remain, which this study tries to tackle. There has been much policy and academic research on the subject in recent years, including, importantly, a major multiyear research project led by the Central Bank of Brazil (CBB). The growing body of research on the subject has provided some understanding of the factors behind the high interest spreads.  A list of potential determinants of high spreads have been identified, including implicit and explicit taxation of financial intermediation, costs and inefficiencies in the banking industry, and risks associated with the contractual environment. There remain significant disagreements, however, regarding methods, results, and policy implications of the existing research. This report tackles the question why intermediation spreads are so high in Brazil, what is the effect and what kind of policies can be implemented in order to reduce spreads. It investigates the various reasons behind the intermediation spreads in Brazil and incorporates analyses from diverse angles. Importantly, it includes a macroeconomic perspective with some innovative conclusions on riskiness of Brazilian public debt. It also provides fresh microeconomic perspectives on the factors that exacerbate the wedge between deposit rates and lending rates. Finally, the study analyzes some of the effects of the high lending rates on the real economy and the corporate sector. 

The main results, messages, and policy implications of the study are summarized below. Macro factors play a first-order role in the determination of spreads, as there is a systematic relationship between the level of the basic interest rate (the SELIC) and the width of the intermediation spreads, magnified by costs, leverage, and default risk. The study submits theoretical arguments and empirical evidence in support of the thesis that the level of the SELIC rate (which is the predominant measure of the opportunity or marginal cost of funds in Brazil) is linked by a non-linear relationship. The corollary is that the story of high intermediation spreads is largely a story of a high SELIC rate. This has three important implications for the analysis and policy debate. 

·  Contrary to the conventional wisdom, Brazil’s high interest margins no longer appear to be unduly mysterious, once account is taken of the high level of the SELIC.

·  The direct impact on spreads of the traditional microeconomic culprits—implicit and explicit financial taxation, operational costs, weaknesses in the contractual environment, etc.—is likely to be of second-order importance at high levels of the SELIC rate but will “bite” increasingly as this rate declines. This is not an argument to relax policy reforms aimed at addressing such culprits.  Such reforms should be vigorously pursued but expectations regarding their effects in the short run have to be tamed.  The largest “bang for the reform buck” in terms of reducing intermediation spreads would come, a this stage, from sustainable reductions in the SELIC rate.

·  To understand the unusually wide intermediation spreads in Brazil we have to understand why the (rate) SELIC rate is so unduly high.

This study opens a number of avenues for future research.  Given the importance of improving the functioning of the financial system for the economy as a whole, the continuation of the work by the CBB on financial development in general should be a high priority.  Specific areas for further study and data collection efforts include the following:

·   Sort out among alternative theoretical formulations of the link between the level of the SELIC to the width of the intermediation spread. 

 

·   Analyze of the interactions between institutional factors and macroeconomic fundamentals in the determination of the high SELIC rate.

 

·   Analyze the predictive power of market expectations, based on the rich data already collected and reported by the CBB.

 

·   Generate a time-series panel data from the Central de Risco (staring with the data for December 2003 used in this study).  

 

·  Ascertain the actual impact on interest spreads of directed lending schemes. 

 

·  Analyze how the marginal and the average cost of funding affect lending spreads for specific credit products. 

 

·   Investigate the reasons why banks hold such a large amount of low-yielding assets and their effect on lending spreads and, in that context, assess the role played in this regard by government regulations, including reserve requirements and directed lending schemes.

 

· Analyze competition issues in a disaggregated way, focusing on specific markets for different financial products. 

 

· Examine in depth the impact of high interest rates and intermediation spreads on growth and the performance of the corporate sector.  In this latter regard, additional Central de Risco time series data would be crucial.